Gangliosides, i.e. sialic acid-containing glycosphingolipids, are well-known components of the mammalian cell surfaces. Furthermore, gangliosides have been found to occur in equilibrium with the corresponding lactones. Thus, GM.sub.3 - and GD.sub.3 -ganglioside [NeuAc.alpha.(2-3)Gal.beta.(1-4)Glc.beta.O-Ceramide and NeuAc.alpha.(2-8)NeuAc.alpha.(2-3)Gal.beta.(1-4)Glc.beta.O-Ceramide] have been demonstrated to form lactones where the sialic acid-derived carbonyl group enters into lactone formation with various hydroxyl groups of the NeuAc and Gal residues. Furthermore, it has been shown that such lactones exist in vivo, e.g. in brain tissue and in the membranes of tumour cells, and are not just artifacts of the isolation procedures (Gross, S. K.; Williams, M. A.; McCluer, R. H. J. Neurochem. 1980, 34, 1351-1361). It has also been shown that GM.sub.3 -lactone is much more immunogenic than is the open form of GM.sub.3 -ganglioside (Nores, G. A.; Dohi, T.; Taniguchi, M.; Hakomori, S.-I. J. Immun. 1987, 139, 3171-3176).
It has also been suggested that GM.sub.3 -lactone is a highly rigid structure with the lactone ring in a chair-like conformation (see scheme 1 below) which would mean that its ability to serve as a recognition site or membrane antigen could be enhanced, compared to its parent structure, which is flexible (Yu, R. K.; Koerner, T. A. W.; Ando, S.; Yohe, H. C.; Prestegaard, J. H. J. Biochem. 1985, 98, 1367-1373). Furthermore, the lactone of GM.sub.3 -ganglioside ("GM.sub.3 -lactone") has been suggested to be a tumour-associated antigen on the cells of an experimental mouse melanoma (Nores, G. et al. cited above). In a comparative immunization with GM.sub.3 -ganglioside and the corresponding lactone, the same authors showed that the lactone is the stronger immunogen, and it was suggested that this compound could be the real immunogen despite being a minor membrane component. Furthermore, the activity of a monoclonal anti-melanoma antibody (M2590) with various cells and liposomes was shown to depend in a threshold, all-or-none, fashion on the concentration of GM.sub.3 -ganglioside in the cell membrane or liposome. Also, the antibody was found to cross-react with GM.sub.3 -lactone.
The ganglioside lactones are unstable at neutral pH while acidic conditions favour the lactone in the equilibrium between GM.sub.3 -ganglioside and GM.sub.3 -lactone. However, since GM.sub.3 -ganglioside is itself acidic since it is an acid-containing glycolipid, the ganglioside might possibly induce its own lactonization when the concentration is sufficiently high in a cell membrane or liposome. This might help to explain the threshold effect described above and might have similar implications for other sialic acid-containing saccharides.
Thus, although a ganglioside lactone has been shown to be far more immunogenic than the open-form ganglioside, the low equilibrium concentration of the lactone at pH-values close to neutral will probably render the lactone to be a rather inefficient immunogen. Thus, it is clear that it would be desirable to have a non-labile immunogenic compound resembling the lactone form of a ganglioside and which is capable of raising antibodies (and other entities of the immune system) capable of at least partial cross-reaction with the ganglioside lactone.